[llvm] [LLVM][CodeGen][AArch64] Don't scalarise v8{f16,bf16} vsetcc operations. (PR #135398)

[Graham Hunter via llvm-commits]

================
@@ -15904,6 +15905,11 @@ SDValue AArch64TargetLowering::LowerVSETCC(SDValue Op,
   if (LHS.getValueType().getVectorElementType().isInteger())
     return Op;
 
+  assert((!Subtarget->hasFullFP16() &&
----------------
huntergr-arm wrote:

nit: Please add an appropriate error string.

https://github.com/llvm/llvm-project/pull/135398